Programmer habits under the microscope: Fridays, night owls, swearing and music
Don't deploy on a Friday. Real programmers work at night. Swearing in comments means messy code. Music ruins concentration. Researchers have tested all four with commit histories, bug trackers and field experiments. Each Question takes one belief, looks at the study behind it, and asks what the study found and how far you can trust it. For developers, team leads, and anyone who has quoted a study in a code review.
Questions
- Not answered. A 2005 paper was titled "Don't Program on Fridays!" What did it actually find?
- Not answered. How did the researchers decide which earlier change caused a later fix?
- Not answered. Which statements does the Eclipse table support?
- Not answered. Does the Eclipse result show that deploying on a Friday causes bugs?
- Not answered. A study published in 2025, "TGIF", tracked commit times in 4,549 GitHub repositories from 2015 to 2024. What did it find?
- Not answered. How many large open source projects did Claes and colleagues study to find out when programmers work?
- Not answered. What did the study find about when most of the active developers commit?
- Not answered. For each project, the researchers found its busiest 8-hour window on weekdays. On average, what share of a project's commits fell inside that window?
- Not answered. In Firefox, the researchers checked whether each top developer was paid by Mozilla or was a volunteer. What did they find?
- Not answered. A commit's timestamp says 23:40. Which of these could make it a poor record of when the work happened?
- Not answered. A 2023 bachelor's thesis at KIT asked whether swearing in code goes together with code quality. How large were the two groups it compared?
- Not answered. Which programming language was every repository in the swearing study written in?
- Not answered. SoftWipe scores code from 0 to 10 for how closely it follows coding standards. What average scores did the two groups get?
- Not answered. Which of these are good reasons not to read the thesis as "swearing makes code better"?
- Not answered. How many software developers took part in Teresa Lesiuk's 2005 field study of music at work?
- Not answered. Lesiuk's developers worked for a week as usual, then two weeks with music, one week without, and one more week with music. Why take the music away in week 4?
- Not answered. Which results did Lesiuk report?
- Not answered. Only one of the four studies changed a habit on purpose and measured what happened. Which one?